Digital Image Processing for Optical Metrology

Abstract

The basic principle of modern optical methods in experimental solid mechanics such as holographic interferometry, speckle metrology, fringe projection, and moiré techniques consists either of a specific structuring of the illumination of the object by incoherent projection of fringe patterns onto the surface under test or by coherent superposition (interference) of light fields representing different states of the object. A common property of the applied methods is that they produce fringe patterns as output. In these intensity fluctuations the quantities of interest such as coordinates, displacements, refractive index, and others are coded in the scale of the fringe period. Consequently one main task to be solved in processing can be defined as the conversion of the fringe pattern into a continuous phase map taking into account the quasisinusoidal character of the intensity distribution. The chapter starts with a discussion of some image processing basics. After that the main techniques for the quantitative evaluation of optical metrology data are presented. Here we start with the physical modeling of the image content and complete the chapter with a short introduction to the basics of digital holography, which is becoming increasingly important for optical imaging and metrology. Section 19.2 deals with the postprocessing of fringe patterns and phase distributions. Here the unwrapping and absolute phase problems as well as the transformation of phase data into displacement values are addressed. Because image processing plays an important role in optical nondestructive testing, finally in Sect. 19.3 some modern approaches for automatic fault detection are described.